AAAAgghr only 16bit

[anyhorizon]

The defense rests.

Peter

[alfspanners]

Ok thanks for the input guys and thanks again Frank for doing your new tests.

I'm worn out with all of this lol, but here goes.

I'll deal with your first test. Frank recorded a 100hz wave via his D8b using Studio One software. In cubase the statistics says 24bit. Ok, first off - why would a d8b file recorded directly in cubase say 16bit in the stats and why does the studio one file say 24bit in the cubase stats.

Well it may have something to do with how it got from studio one to cubase in the first place. I tried exporting a mackie file that was reading 16bit in the stats, then reimporting and found that cubase was saying the new file was 24bits. I phase reversed the new file and exported the seemingly silent leftover from the cancellation. When I added 40db of gain to it and zoomed in - this was what i got.

cubase export small.jpg (Array KiB) Viewed 1565 times

Notice the very small occasional 24bit blips. These have clearly been added on export and allthough I'm no expert I would hazard a guess that this is caused by tiny calculation errors in the mixdown. This may just be cubase, but i have heard that digital audio processing is not as exact as some people may think, especially on something that has gone through a mix bus - I mean why has there been the clamour to go 64bit?
These 24bit step sized blips would indicate to cubase that the file was 24bit.

Of course your desk may be 24bit, I just throw that one in because we need to be certain that we are not corrupting those files by our tests.

Secondly, your image of bitter looks very like the image I had - notice the two very wide bits on 17 and 18. According to the bitter manual this suggests they are stuck bits (no i don't really know what that means, but it doesn't sound good and, much more importantly - why do I only get these stuck bits with d8b files, why does bitter never show any stuck bits on my home files recorded on a fireface? If bitter is malfunctioning why does it only do it with d8b files? there must be some answer).

Frank's next file shows the white noise generated in Amadeus pro. Cubase stats estmate 20/21bit, bitter show 24bits.The 2 results are not inconsistant. Bitter is showing the number of bits in use at any time, If I'm correct Cubase isn't looking at the how many bits are being used ie 0s and 1s flashing on and off - it's analysing what the smallest step size of the constructed wave is and arriving at it's estimation this way. Remember white noise is random, it may take some time for two comparable samples with a 24bit step size to be generated.

The problem I have with the rest of Frank's examples is Bitter Correct and Cubase stats wrong, are they both wrong, Frank stated that Cubase was full of shit - is bitter right then? How do you explain the stuck bits?

i WANT TO MAKE THIS PERFECTLY CLEAR -nO ONE HAS COME UP WITH AN EXPLANATION ABOUT MY BLOWNUP WAVEFILE COMPARISONS BETWEEN MACKIE FILES AND KNOWN 24BIT FILES. Just to demonstrate further I downloaded audacity, which doesnt vertaclly zoom so I had to add 80db to the files. Also because they are extremely boosted the db scale compacts them the greater the amplitude so only compare like amplitudes.

D8b file 80db boost

audacity mackie mag small.jpg (Array KiB) Viewed 1565 times

Fireface file 80db boost

Audacity home mag small.jpg (Array KiB) Viewed 1565 times

Clearly the Rme has smaller step sizes. Incidentally the Mackie file I used hasn't been anywhere near cubase - it was done on an Hdr 24/96 years ago.


IU'm really sorry if I'm annoying people and you don't think very much of my methods, I came to this forum, not to prove a point, nor for an ego joust - I just wanted to know if other v3 users had the problem or it was an issue with our d8b or if v5.1 fixed it. I've tried to behave respectfully to all forum members and thanked them for taking the time to do tests.

We have decided to replace the desk, based on my recommendation so when my methods are described as flawed...well it's my neck on the line so I have to defend my position by testing that assertion - I'm not doing it to score points or wind folks up.

Anyway - this is exhausting and very stressfull.

Mark

[anyhorizon]

We have decided to replace the desk, based on my recommendation so when my methods are described as flawed...well it's my neck on the line so I have to defend my position by testing that assertion

I think you'll regret it but good luck anyway. I assume you never tried the RME DigiCheck bitscope.

Peter

[FrankH]

Well it may have something to do with how it got from studio one to cubase in the first place.

D8B recorded directly into Studio One. Resulting waveform dragged and dropped into the Cubase Timeline. No Exportation used.

I assume you never tried the RME DigiCheck bitscope.

What he said.

And finally: When the crapstorm erupted over the "16bit Aux Out" debacle, years ago, the geeks of the old forum population went over the board with a fine tooth comb, looking for every possibility of "less than 24bits" on every I/O pathway the D8B offers. They were on a mission. Some of the rabble even had the temerity to use real-world digital audio test gear like Audio Precision APx585's and the like to check the throughput. Funny how this didn't show up in any of the findings.

[High C Double G]

24 bits? I don't even have 24 horses! (whinny)

[anyhorizon]

"Hello, I'm Mr. Ed"

[hyaena]

Hello Mark, sorry, it took some time to gather the facts.

So, first of all, I don´t have a clue what you guys are talking about in bits and pieces. My understanding of digital isn´t sophisticated at all, but I still think I got most of it.

I tried to get the fellas from the measuring technique at my radio-station involved. But they´re busy with anything else than watching quality. They don´t have time to think about what they´re doing.

I have run some tests with my equipment into RME RayDat and the DigiCheck measuring tool.
So here are the results, that keep me confused. So sorry, fellas, I´m not able to analyze the pictures, I just put them here for discussion.

First, a recording done into the mic-input 1 of my d8b at gain -22dB, input 1 + 2 being tested by DigiCheck:

Second, into the mic-input 1 of the d8b at unity gain, input 2 no modulation, input 1 + 2 being tested by DigiCheck:

Third, into the mic-input 1 of the d8 at unity gain, mic being talked at so the mod L will be resulting in -60dB level - no Mod R, input 1 + 2 being tested by DigiCheck:

Next, GoldenAge PreAmp into Symetrix 620, 20bit - no mod L - Mod R, input 35+36 of the RayDat being tested by DigiCheck - what makes me wonder, because the tool is supposed to be 20bit only.:

5th and last DIGI-Check Behringer ADA8000 24bit, mod L - no Mod R, , input 25+26 of the RayDat being tested by DigiCheck:

I have some difficulties uploading the screenshots, sorry.
But you guys seem to have your opinion already, so I go to sleep now.

Manfred

[anyhorizon]

Uh, where are the results?

Peter

[Dan Worley]

Sample-zoomed diagrams don't make for good listening.

RME's Digicheck works great.

Another good tool is SSL's X-ISM. It's free. Scroll down to the bottom of this page (registration is required). It's not really made for measuring device outputs, it's used for interpreting inter-sample peaks, but I have found its bitscope to be quite useful and telling.

I tested my d8b about 100 times when I was using it.

Good luck.

c-ya,

Dan Worley

[Dan Worley]

Uh, where are the results?

Peter

On the grassy knoll.

c-ya,

Dan